Download SPRA972 - Texas Instruments

Application Report
SPRA972 − November 2003
TMS320VC5510 Power Consumption Summary
Ryan Verret
C5000 Applications
ABSTRACT
This document assists in the estimation of power consumption for the TMS320C5510 digital
signal processor (DSP). As power consumption can vary widely on this device, a spreadsheet
was developed to provide a better estimate. This allows the user to tailor the prediction to their
particular application. It also allows designers the ability to test the efficiency of different
configurations before any hardware is assembled or any code is written.
Contents
1
Activity-Based Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1 Static Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2 Activity Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3 Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2
2
2
3
2
Using the Power Estimation Spreadsheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1 Choosing Appropriate Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.1 Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.2 Frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.3 Idle Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.4 % Utilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.5 % Writes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.6 Bits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.7 % Switch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.8 Trace Length . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.9 Load Capacitance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.10 Other . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2 Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3 Graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3
3
3
3
4
4
5
5
5
5
5
5
5
6
3
Using the Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
4
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Trademarks are the property of their respective owners.
1
SPRA972
1
Activity-Based Models
Power consumption of the TMS320C5510 DSP is extremely application dependent due to its
many peripherals and variety of power saving modes. It is because of this that the application to
be measured must be well understood before a power estimate can be given. The activity of
each peripheral must be known in order to configure the power estimation spreadsheet for
accurate results which can be used in power supply design or battery life prediction.
The model used in this spreadsheet is based on two modes of power consumption: static and
activity power. With this model, each active component can be isolated and accurately modeled
to determine its contribution to the static power.
1.1
Static Power
Static power is the power consumption inherent to the CMOS technology itself (sometimes
called leakage power). Static power is not affected by the device activity or operating frequency.
This power component is measured when the PLL is not enabled and nothing on the device is
being clocked. Static power is affected by core voltage, IO voltage, and the device operating
temperature.
1.2
Activity Power
Activity power is the consumption of the active parts of the DSP. These include the CPU, EMIF,
PLL, peripherals, etc. Power consumption is based on voltage, frequency and the given
configuration of each module. In order to provide more accurate power estimation, each block
can be characterized independently and its relative contribution to overall consumption is
provided. This aids in system design for greater efficiency.
Each module has parameters used to describe its activity. These include frequency, idle status,
utilization, read/write balance, bus size and switching probability, trace length, and load
capacitance. Each module may not include all of these parameters, however.
2
•
Frequency is the operating frequency of a module, or the operating frequency of the
interface to that module.
•
Idle Status indicates whether the module is idle or active.
•
%Utilization is the percentage of activity in a module relative to its maximum.
•
%Writes is the percentage of writes relative to the total number of transfers (writes and
reads).
•
Bits is the number of data bits in use on a interface that supports variable-width busses.
•
% Switch is the probability that a data bit will switch from one cycle to the next.
•
Trace Length is the total board trace length being driven by each pin of the interface.
•
Load Capacitance is the sum of the input capacitances at the end of each trace.
TMS320VC5510 Power Consumption Summary
SPRA972
1.3
Modules
Each of the following modules is user-configurable in the C5510 power estimation spreadsheet
within realistic operating parameters.
•
•
•
•
•
•
•
•
•
•
•
•
•
2
DPLL
CPU
CLKOUT
EMIF
DMA (6 channels)
HPI (as part of DMA)
Instruction Cache
McBSP0
McBSP1
McBSP2
Timer0
Timer1
GPIO
Using the Power Estimation Spreadsheet
To use the spreadsheet, you simply enter the usage parameters into the white cells. To ensure
that the data validation feature limits the input to realizable configurations, enter values from left
to right, top to bottom. The spreadsheet will then take the provided information and display the
details of power consumption for that configuration.
To simplify its operation, the spreadsheet does not support more than one idle configuration at
any given time. For applications where the idle status is constantly changing, each possible
configuration should be estimated independently, then the maximum taken for power supply
design or all cases time-averaged to predict battery life.
2.1
Choosing Appropriate Values
It is critical to choose the correct values for the power estimation spreadsheet, to produce
accurate results. Each module must be considered in isolation and you must account for all
activity included in a given operation. For instance, ‘EMIF current’ is not included in the
instruction cache ‘miss current.’ This must be included separately in the EMIF section.
2.1.1
Temperature
The static power is affected by the operating temperature of the device. Select the desired
temperature (in degrees Celsius) by entering the temperature in the “Temp” cell at the top of the
spreadsheet.
2.1.2
Frequency
The module frequency should be straightforward to determine. It will either be the DPLL
frequency, or some integer fraction of it. For some modules, such as the timers, the frequency is
automatically set to the DPLL frequency for core power consumption and set to the specified
frequency for IO power consumption.
TMS320VC5510 Power Consumption Summary
3
SPRA972
2.1.3
Idle Status
Idle status for a given module is simply whether or not that module is configured by software to
be in its idle state. The spreadsheet only supports on idle configuration at any time.
2.1.4
% Utilization
Utilization is explicitly defined for each module in order to provide a more accurate estimate of
power consumption. If a module is not listed, then it is assumed to be in use whenever it is not
idle.
4
•
CPU – Since there are varying degrees of activity for the CPU, it is more difficult to provide a
utilization number. Whenever the CPU is active (non-idle) it is repeatedly executing
instructions. For this reason, 0% activity will be defined as a repeated NOP instruction– the
smallest amount of power the CPU can consume while active. Conversely, 100% activity will
be defined as the most power intensive instruction – the dual multiply and accumulate. All
other instructions will fall somewhere in between. No single algorithm will achieve 100%
utilization, but some highly optimized functions can come close. In addition, the CPU must
also be used for control oriented tasks that consume far less current.
For example, assume that a certain application executes control code half of the time and a
highly optimized algorithm for the other half. If the control code is estimated to be at 30%
utilization and the dense DSP code is estimated to be at 90% utilization, the overall
utilization would be 60% (30% × 50% + 90% × 50%). If the application spent more time
executing the optimized algorithm, utilization would obviously go up, and vice versa. By
examining individual portions of an application and estimating the utilization and time spent
in each, a more accurate CPU utilization percentage can be obtained.
•
CLKOUT – The clock output utilization percentage is simply defined as the percentage of
time that the DSP core clock is being output. On the 5510 design, the CLKOUT output buffer
is operated by the CVdd supply (the spreadsheet credits power consumed by the CLKOUT
correctly).
•
EMIF – EMIF utilization is related to the maximum bandwidth of the EMIF. One hundred
percent utilization corresponds to the maximum transfer rate for a given frequency. This
number will be scaled down by both slower and less frequent transfers. As the minimum
period for EMIF accesses is three cycles, the utilization percentage is defined as three times
the average frequency of activity over the entire block of code being analyzed divided by the
actual EMIF clock frequency. The average activity frequency is simply the average number
of EMIF accesses per second and the EMIF clock frequency is some fraction of the DPLL
frequency.
Choose the desired memory type (synchronous or asynchronous) in the “Other” column.
•
HPI – HPI utilization is also related to the maximum bandwidth of the HPI. The minimum
period for HPI accesses is 14 CPU cycles + 10ns (as referenced in the 5510 data manual,
literature number SPRS076) At 200 MHz, this is equivalent to 16 cycles, so the utilization
percentage is defined as sixteen times the average activity frequency divided by the DPLL
frequency.
•
DMA – DMA utilization for a given channel is related to the maximum bandwidth of the port
for each data resource, which is one transfer each cycle. This results in a utilization number
that is simply the DMA channel’s average activity frequency divided by the DPLL frequency.
•
Instruction Cache – Utilization of the instruction cache is defined as the percentage of time
that the CPU executes instructions from external memory.
TMS320VC5510 Power Consumption Summary
SPRA972
•
McBSP – McBSP utilization is defined as the percentage of time that the McBSP is
transferring data.
•
Timer – Timer utilization is defined as the percentage of time that the timer is counting. The
timer can be configured to use the TIN/TOUT pin as an output, an input or not used. If this
pin is used as an output, select TOUT enabled in the “Other” column. Otherwise, select
TOUT disabled.
•
GPIO/XF – Utilization for general purpose outputs is the percentage of time that they are
switching at their specified frequency. Select the number of GPIO pins used in the “Other”
column of the spreadsheet.
2.1.5
% Writes
For modules that move data onto a bus, % writes is defined as the percentage of utilization in
which the peripheral is putting data on the bus. Data is assumed to be taken off the bus for the
remainder of the utilization period. For modules with equal input / output bandwidth, the write
percentage should be 50%. And for the instruction cache, % writes is the hit percentage of
instruction accesses.
2.1.6
Bits
Bits is the number of data bits on a variable width bus. The spreadsheet only permits valid bus
sizes.
2.1.7
% Switch
Random data has a 50% chance that any bit will change from one cycle to the next. Some
applications may be able to predict this change using some a priori information about the data
set. If there is a property of the algorithm that allows prediction of the bit changes, the
application-specific probability can be used. All other applications should use the default number
of 50%.
2.1.8
Trace Length
The trace length parameter indicates the average length of controlled-impedance board trace of
each output for a given interface (assumed 50Ω trace). This only affects IO power and further,
only contributes when the modules are writing.
2.1.9
Load Capacitance
Load capacitance is the average of the sum of the input capacitances of all external devices
connected to a given module’s outputs. This parameter should be obtained from the datasheets
of the devices that are connected to the DSP.
2.1.10
Other
Some modules have other parameters that are self-explanatory. These include whether the
EMIF is configured in synchronous or asynchronous mode, if the timers are configured as
outputs, and if the number of GPIO pins configured as outputs.
2.2
Units
The results are estimated in the spreadsheet and displayed in milliamps (mA) or millliwatts
(mW). Click on the units in the “Total” row of the calculated results and use the pull-down menu
to select the desired units.
TMS320VC5510 Power Consumption Summary
5
SPRA972
2.3
Graphs
The graphs included in the spreadsheet show the relative contribution of total core and IO power
for each module. They provide a visual display of power usage and allow easy identification of
the major power consumers.
3
Using the Results
The results presented in the spreadsheet are based on measured data for revision 2.2 silicon.
The measured units were selected to be towards the high end of power consumption for
production units. Most production units will have power consumption that is below the value
given in the spreadsheet, if accurately modeled.
The spreadsheet produces an estimated average current (or power) for the specified
configuration. Transient currents may cause power to exceed the estimated value for short
periods, but long term power consumption should be at or below the spreadsheet value. This
allows for better estimates of power supply requirements and more accurate battery life
predictions.
4
References
1. TMS320VC5510 Fixed-Point Digital Signal Process Data Manual (SPRS076)
6
TMS320VC5510 Power Consumption Summary
IMPORTANT NOTICE
Texas Instruments Incorporated and its subsidiaries (TI) reserve the right to make corrections, modifications,
enhancements, improvements, and other changes to its products and services at any time and to discontinue
any product or service without notice. Customers should obtain the latest relevant information before placing
orders and should verify that such information is current and complete. All products are sold subject to TI’s terms
and conditions of sale supplied at the time of order acknowledgment.
TI warrants performance of its hardware products to the specifications applicable at the time of sale in
accordance with TI’s standard warranty. Testing and other quality control techniques are used to the extent TI
deems necessary to support this warranty. Except where mandated by government requirements, testing of all
parameters of each product is not necessarily performed.
TI assumes no liability for applications assistance or customer product design. Customers are responsible for
their products and applications using TI components. To minimize the risks associated with customer products
and applications, customers should provide adequate design and operating safeguards.
TI does not warrant or represent that any license, either express or implied, is granted under any TI patent right,
copyright, mask work right, or other TI intellectual property right relating to any combination, machine, or process
in which TI products or services are used. Information published by TI regarding third-party products or services
does not constitute a license from TI to use such products or services or a warranty or endorsement thereof.
Use of such information may require a license from a third party under the patents or other intellectual property
of the third party, or a license from TI under the patents or other intellectual property of TI.
Reproduction of information in TI data books or data sheets is permissible only if reproduction is without
alteration and is accompanied by all associated warranties, conditions, limitations, and notices. Reproduction
of this information with alteration is an unfair and deceptive business practice. TI is not responsible or liable for
such altered documentation.
Resale of TI products or services with statements different from or beyond the parameters stated by TI for that
product or service voids all express and any implied warranties for the associated TI product or service and
is an unfair and deceptive business practice. TI is not responsible or liable for any such statements.
Following are URLs where you can obtain information on other Texas Instruments products and application
solutions:
Products
Amplifiers
Applications
amplifier.ti.com
Audio
www.ti.com/audio
Data Converters
dataconverter.ti.com
Automotive
www.ti.com/automotive
DSP
dsp.ti.com
Broadband
www.ti.com/broadband
Interface
interface.ti.com
Digital Control
www.ti.com/digitalcontrol
Logic
logic.ti.com
Military
www.ti.com/military
Power Mgmt
power.ti.com
Optical Networking
www.ti.com/opticalnetwork
Microcontrollers
microcontroller.ti.com
Security
www.ti.com/security
Telephony
www.ti.com/telephony
Video & Imaging
www.ti.com/video
Wireless
www.ti.com/wireless
Mailing Address:
Texas Instruments
Post Office Box 655303 Dallas, Texas 75265
Copyright  2003, Texas Instruments Incorporated